In video game development, it can be challenging to predict the emotional responses players may have to newly released video games. While prior research has examined emotions in games, few studies have addressed how diverse gameplay structures shape emotional expression. Understanding players' emotions and experiences can help developers and researchers enhance gameplay. Therefore, this study investigates the sentiment and length of player reviews on the Steam platform across ten player engagement modes (PEMs) such as single-player, cooperative, and massively multiplayer games, among others. A dataset of 15,032,567 textual reviews was collected using Steam's API, focusing on the most recent 1000 reviews per game. The analysis explored how PEMs affect review length, playtime at review, and emotional expressions. The results show that significant differences exist in player behavior and sentiment between engagement modes. Negative reviews were longer and written earlier than positive ones. Furthermore, sentiment analysis using the NRC Emotion Lexicon identified various differing emotions between PEMs. The results can help developers better understand the emotions that players associate with the different PEMs, while they also show that early player experiences should be prioritized. Additionally, the results offer researchers a foundation for future studies regarding players' emotions and experience in video games with various PEMs.
Guzsvinecz et al. (Tue,) studied this question.